CHAPTER 3

A BASOLATERAL K⁺ CONDUCTANCE DOMINATES THE RESTING MEMBRANE POTENTIAL OF SMALL INTESTINAL CRYPTS.

A basolaterally located K⁺ conductance is a characteristic feature of fluid transporting epithelia and has been shown to be important in the regulation of fluid and electrolyte transport, in cell volume regulation and in the maintenance of cell membrane potential. Such a K⁺ conductance is thought to play a role in the recycling of K⁺ ions taken up into the cell by the Na⁺-K⁺ ATPase and by the NaK2Cl cotransporter in the unstimulated cell, following Na⁺-coupled nutrient absorption and also during Cl^- secretion. The hyperpolarisation of cell membrane potential by a basolateral K⁺ conductance maintains an electrochemical gradient favourable to Cl^- exit from and Na⁺ entry into the cell (reviewed by Dawson & Richards, 1990). Investigators have employed a variety of approaches to establish the presence of a preferentially K⁺-permeable basolateral membrane in epithelia, including conventional microelectrode techniques (Valverde et al., 1991), measurements of ⁸⁶Rb⁺ efflux (Mandel et al., 1986b) and the use of established K⁺ channel blockers such as Ba²⁺ and TEA (Burckhardt & Gogelein, 1991).

The suggestion that such a basolateral K⁺ permeability might be present in the mammalian small intestinal epithelium was supported by studies of unstimulated monolayers of the model intestinal T84 cell line, where the rate of ⁸⁶Rb⁺ efflux across the basolateral membrane has been shown to be ten times the rate of efflux across the apical membrane (Mandel et al., 1986b). Later electrophysiological measurements from the small intestinal epithelium of Necturus indicated that the basolateral membrane is dominated by a K⁺ conductance (Valverde et al., 1991). Studies of ionic permeabilities in mammalian enterocytes, however, have thus far been largely confined to dissociated single cell preparations, thereby losing epithelial polarity. Patch-clamp studies of dissociated preparations of guinea-pig villus enterocytes have revealed that whole-cell K⁺ currents are dominated by a Ca²⁺-independent inwardly rectifying K⁺ conductance, although it is not known in which membrane domain such a K⁺ conductance resides (Sepulveda et al., 1991).

The presence of a basolateral membrane selectively permeable to K⁺ has yet to be demonstrated in the mammalian small intestinal epithelium. In order to investigate whether such a conductance is present we have used the perforated-patch recording technique to study the ionic permeabilities present in the unstimulated small intestinal crypt. The results of these experiments employing ion substitutions and K⁺ channel inhibitors indicate that the basolateral membrane of unstimulated small intestinal crypts is dominated by a quinine- and Ba²⁺-sensitive K⁺ conductance.

The initial stage of preparation of intact crypts was performed as described in section 2.2.2.. Fractions were collected over timed 10, 6, 2, 2 and 2 minute intervals. The three fractions from minutes 17-22 were centrifuged at 50g for 1 minute in a Super Minor MSE centrifuge and washed in Hanks medium at 4^oC. The fractions were then centrifuged at 50g, the supernatent was aspirated and the pellets were pooled by resuspension in 25ml of DMEM (pH 7.4 with 0.1M NaOH) containing 2mM DTT at 4° C. The crypt suspension was then placed on a rotary inversion mixer for 20 minutes. The suspension was then spun at 50g for 2 minutes and aspirated, before final resuspension in 2ml of ice-cold DMEM. The cells were maintained on ice and used within 6 hours. The crypts were easily identified under phase contrast microscopy by their birefringence and cylindrical morphology and appeared viable as judged by their capacity to exclude 0.03% trypan blue.

The composition of the intra- and extracellular solutions is given in table 3.1. Patch pipettes were filled with an artificial K⁺-rich "intracellular" solution containing 100 m g/ml of nystatin. Nystatin-containing pipette solutions were prepared freshly every 1-2 hours (from a frozen stock (10 mg/ml) of nystatin in dimethylsulphoxide (DMSO)) by adding 20m l of nystatin stock to 1ml of pipette solution in an Eppendorf tube, taking precautions to avoid exposure to light. The tube was sonicated for 2 minutes and the pipette solution was then filtered through a Dynaguard (0.2 m m) ME syringe filter (Microgon inc., U.S.A.) into a foil-covered 5ml beaker kept on ice. In some experiments pipette Ca²⁺ was buffered to a quasiphysiological intracellular concentration [93 nM] with EGTA. The spontaneous "zero-current" reversal potential (E_m) was not affected by this manoeuvre, which can be explained by the impermeability of nystatin to divalent cations (Horn and Marty, 1988). The concentrations of Ca²⁺ [2 mM] and EGTA [5 mM] required were calculated according to the stability constants of Martell & Smith (1974). The tip of the pipette was filled by capillary action with nystatin-free pipette solution before the pipette was back-filled with the nystatin-containing solution. The composition of the pipette solution is as given in table 3.1.

The low Na⁺ solution used for extracellular ion replacement experiments was prepared by equimolar replacement of NaCl with N-methyl-D-glucamine Cl and the low Cl^- solution by an equimolar replacement with Na gluconate. Changes in [K⁺]_o were achieved by equimolar substitutions of KCl and NaCl. Hanks medium containing either 5mM TEA or 5mM Ba²⁺ was prepared omitting 5mM NaCl. All bath and pipette solutions were filtered through Dynaguard 0.2m m syringe filters prior to use. Osmolarities of solutions were checked using a freezing-point depression osmometer (3MO Advanced Instruments, Massachussetts, USA) and was adjusted to 295 mosmol/l by the addition of D-mannitol if necessary.

3.2.3. Perfusion arrangements for crypt electrophysiological recordings.

A suspension of isolated crypts was aliquoted onto type 0 glass 10mm diameter cover slips (Chance Propper) precoated with 0.01% polyethylene imine. The pretreated cover slips were supported on 2.5cm glass micro-fibre filters (Whatman Ltd., England) and placed inside a 135mm petri dish. The crypts were left to equilibrate for 10-15 minutes at room temperature before being transferred into a purpose-built chamber containing Hanks medium, the bottom of which comprised of a type 0 glass cover slip. The chamber was mounted on the stage of an inverted microscope (Zeiss IM 35, Germany). Rapid solution changes were effected by directing a small jet of the desired solution directly onto the crypt under study, without substantially changing the composition of the bulk solution, by means of a perfusion loop similar to that described by Suzuki et al (1990). The perfusion arrangements for electrophysiological experiments were as shown in Fig.3.1.. The bulk solution was constantly removed with aid of a peristaltic pump and replenished by Hanks medium flowing under gravity. All experiments were conducted at room temperature (22-25° C).

Patch-pipettes were fabricated from Boralex glass capillaries of outside diameter 1.7mm (Rochester Scientific Co., New York, U.S.A.) using a two-stage vertical pipette puller (PP-83, Narishige, Japan). The tips of pipettes were polished using a microforge before back-filling with "intracellular" solution. Patch pipettes used for current-clamp recordings had resistances of 4-5 MW when filled with KCl-rich solutions. "Zero-current" reversal potentials and whole-cell currents were recorded according to the method of Hamill et al. (1981) with a List EPC-7 patch-clamp amplifier (List Electromedical, Germany) using the perforated-patch method as previously outlined by Horn & Marty (1988). Isolated crypts attached to 10 mm diameter type 0 glass cover-slips (Chance-Propper) were mounted in a specifically designed perfusion-chamber, as described above, and the crypts were viewed on the stage of an inverted microscope (Zeiss IM 35, Germany) at a total magnification of x320.

The smooth basolateral membrane of cells in the mid-crypt region was approached with patch-pipettes without applying positive pressure to the pipette using a hydraulic micromanipulator (Narishige MO 203, Japan). Giga-ohm seals were obtained by applying light-suction to polished patch-pipettes gently pressed against the crypt membrane (see Fig.1.4.). During selected recordings a train of small voltage-pulses (0.1-10 mV) was applied to the stimulus input of the amplifier to monitor the progression of seal formation by measuring the size of the resulting current pulses. Access resistances were estimated to be in the range of 20-50 MW within 1-3 minutes of seal formation. For whole-cell current recordings the series conductance was adjusted in parallel with the series capacitance to compensate for the whole-cell capacitative current. Access resistances were monitored and progressively compensated throughout voltage-clamp recordings. Cells were voltage-clamped at a holding potential (typically -40 mV) and membrane currents were recorded in response to both depolarising and hyperpolarising voltage steps.

In current-clamp experiments the zero-current potential of the crypt was continuously monitored to provide a measurement of membrane potential (E_m). Stable E_m values had been obtained, usually within 1-3 minutes of seal formation, indicating that the pipette-filling solution had equilibrated with the cytoplasm. Membrane potential changes evoked by ion substitutions and by the addition of channel inhibitors to the bathing medium were recorded, with or without the application of current pulse trains of fixed amplitude to monitor changes in membrane conductance. The zero-current reversal potentials are referred to as membrane potential values (or E_m) throughout the text, assuming that junction potentials after seal formation in low access resistance perforated-patch recordings are 3/45mV (Rae & Fernandez, 1991).

The established sign convention is used throughout and potentials are reported with respect to the patch-pipette. The bath electrode consisted of an Ag-AgCl pellet. Junction potentials evoked following bath solution changes were determined and measurements obtained from recordings were corrected for the measured offsets.

An IBM-AT microcomputer equipped with a Lab-PC laboratory interface (National Instruments, U.S.A.) and software programmes (VGEN and VCAN, J. Dempster, Dept. of Physiology and Pharmacology, University of Strathclyde, Glasgow) were used for data acquisition and the analysis of current-clamp and voltage-clamp recordings. Voltage pulse protocols and current pulse trains were applied to the stimulus input of the List EPC-7 following digital-to-analogue conversion using the Lab-PC laboratory interface. The signal from the patch-clamp amplifier was simultaneously viewed on a storage oscilloscope (Gould 1421) and recorded on videotape for subsequent analysis together with triggering pulses using an adapted pulse-code modulation encoder (Sony PCM-701ES) as described by Lamb (1985).

Replayed records of whole-cell currents were low pass filtered at 2.5kHz (-3db) using a variable 8-pole Bessel filter (Barr & Stroud) and then digitised using a PC-Lab laboratory interface. Current-clamp recordings were acquired at a frequency of 4 Hz and whole-cell recordings at 2-5kHz. Results are expressed as means ± standard errors of n observations.

After GÛ seals had been obtained with nystatin-filled micropipettes the membrane potential (E_m) was monitored in the current-clamp mode of the patch-clamp amplifier, and normally reached a stable value within 2-5 minutes of seal formation. The mean value for E_m obtained from the mid-region of the unstimulated crypt was -49 mV ± 2 (mean ± SEM, n = 35) and this remained stable for up to 60 min. In order to test the dependence of E_m upon extracellular ion concentration, ion substitution experiments were performed. Fig.3.2A shows a recording obtained with a crypt originally bathed in normal Hanks solution ([K⁺]=6.2 mM) with an initial E_m value of -57 mV. Increasing the extracellular K⁺ concentration first to 20 and then to 40 mM (equimolar replacement of NaCl with KCl) led to rapid depolarisations of E_m that were readily reversible. Decreasing extracellular K⁺ to 0.5 mM produced a hyperpolarisation of crypt E_m. Replacement of all but 8 mM Cl^- in the bathing medium for gluconate evoked a small hyperpolarisation whilst replacing Na⁺ with N-methyl-D-glucamine was virtually without effect. Fig.3.2B is a semi-logarithmic plot showing mean E_m values as a function of ion concentration. Crypt E_m was strongly dependent on extracellular K⁺, especially at high concentrations where the slope gave a 37 mV change in E_m for a ten-fold change in extracellular K⁺ concentration. The experimentally obtained curve for K⁺ (solid line) can be compared to the Nernstian slope for an ideal K⁺ conductance (broken line), calculated assuming an intracellular K⁺ concentration of 129 mM (taken from De Castillo, 1987). The comparatively small effect upon E_m of Cl^- replacement and the negligible effect of Na⁺ replacement are also summarised in Fig.3.2B.

Effect of ion replacements on the membrane potential of the mid-region of isolated small intestinal crypts. A. E_m recording of an isolated crypt initially bathed in normal Hanks medium of the following composition (mM): 140 NaCl, 6 KCl, 4 NaHCO₃, 1.3 CaCl₂, 0.5 MgCl₂ and 10 HEPES pH 7.4. During the periods indicated by the horizontal bars, the K⁺ or Cl^- concentration was changed to the mM concentration given (equimolar replacement with Na⁺ or gluconate respectively). The initial value of E_m is given. B. Summary of ion replacement experiments with Na⁺, K⁺ and Cl^- ion substitutions (K⁺ and Cl^- substitutions are as above, Na⁺ was substituted with N-methyl D-glucamine). Results are means of E_m values ± SEM from 6-15 experiments.

Experiments were performed to characterise the nature and voltage-dependence of whole-cell currents in unstimulated crypt enterocytes. Figure 3.3A shows the whole-cell currents recorded with a KCl-rich pipette solution at a holding potential of -40mV from the mid-region of an unstimulated crypt enterocyte. Whole-cell current recordings using nystatin-perforated patches could only be obtained from crypts when a low access resistance (20-30 MÛ) patch had been obtained, and the capacitance compensation measurement did not exceed 100 pF. The currents recorded from the mid-region of 3 individual unstimulated crypts appear to show a marked outward rectification (see Fig.3.3D). Furthermore, an apparent time-dependent activation of outward current was observed at more depolarised holding potentials.

In order to establish further the nature of the basolateral K⁺ conductance present in the unstimulated crypt, classical K⁺ channel blockers were added to the bathing medium. Depolarisations evoked by the addition of the K⁺ channel blockers TEA⁺, Ba²⁺ and quinine have been reported in the Necturus gallbladder epithelium (Segal & Reuss, 1990) and rat colonic crypts (Burckhardt & Gogelein, 1991) and have been used to characterise the K⁺ conductance(s) responsible for determining the resting E_m in these epithelia.

The effect upon crypt E_m of addition of 400m M quinine is shown in Fig.3.4A, where a rapid and reversible 59mV depolarisation was observed. In 4 experiments 500m M quinine evoked a 39.8 ± 2.5mV depolarisation (mean ±SEM), which was maximal within about 30s of addition of the inhibitor. The dose-dependency of the effect of quinine upon crypt E_m was determined (shown in Fig.3.4B). At low concentrations (1-10m M) quinine evoked a small (3.3 ± 0.8mV, P<0.05, n=4) hyperpolarisation, whilst at concentrations >50m M quinine evoked a dose-dependent depolarisation of crypt E_m with an IC₅₀ of around 200m M. The effects of 5mM Ba²⁺ upon crypt E_m is shown in Fig.3.4C; 5mM Ba²⁺ induced a slow 17mV depolarisation of E_m, associated with a decrease in crypt membrane conductance. A slow Ba²⁺ blockade (maximal within around 60s in 4 experiments) would also be consistent with channel blockade at the intracellular face if Ba²⁺ can permeate the cell membrane. In 4 experiments 5mM Ba²⁺ induced a mean depolarisation of 16.8 ± 2.0mV which was slow to recover upon washout. The effect of the K⁺ channel blocker tetraethylammonium (TEA) upon E_m was also tested. In 3 experiments 5mM TEA had no significant effect (P<0.05) upon crypt E_m (3 ± 1mV depolarised). The effects of the 3 inhibitors tested is summarised in Fig.3.4D.

The results presented here demonstrate that a basolateral K⁺ conductance, sensitive to Ba²⁺ and quinine, but not to TEA⁺, determines, at least in part, the resting membrane potential of unstimulated small intestinal crypts. The basolateral location of such a K⁺ conductance would be necessary to account for the rapid and reversible effects of K⁺ ion substitutions and channel blockers upon crypt membrane potential, as the narrow lumen of the crypt would be relatively inaccessible to changes in the composition of the bathing medium. Increasing extracellular K⁺ depolarised the crypts, with a 37 mV change in E_m per ten-fold change in extracellular concentration at high [K⁺]_o (Fig.3.2), less than the 58mV per decade change predicted for a purely K⁺ selective membrane conductance by the Nernst relation.

The absence of any effect upon E_m of Na⁺ replacement experiments suggests that a basolateral Na⁺ conductance does not contribute to the resting membrane potential, an argument further strengthened by the absence of an effect of 1mM amiloride upon crypt E_m when added to the bathing medium (result not shown). Similarly Cl^- substitution experiments are inconsistent with the presence of a basolateral Cl^- conductance as no depolarisations were evoked in response to gluconate substitution. The small hyperpolarisation of E_m evoked upon Cl^- replacement, not exclusively due to a junction potential, might be explained by a decreased uptake of K⁺ through a putative cotransport system and a resultant change in E_K. The ion substitution data indicates that other apically-located conductances may be present in the mid-region of the unstimulated crypt. Some indirect evidence for this assertion comes from the effect of the Cl^- channel blocker 5-nitro-2-(3-phenylpropylamino)-benzoate (NPPB) when added to the bathing medium, which evokes a slow hyperpolarisation of crypt E_m (Fig.5.8) consistent with the presence of an apically-located Cl^- conductance. The decrease in the dependence of crypt E_m upon [K⁺]_o at low [K⁺]_o concentrations suggests that a Na⁺ and/or a Cl^- conductance may be activated, possibly as a consequence of the resultant hyperpolarisation.

The rapid and reversible depolarisations evoked by Ba²⁺ and quinine are also consistent with the presence of a K⁺ conductance in the crypt basolateral membrane. However the addition of 5mM TEA⁺ had little or no effect upon E_m in the unstimulated crypt. Several previous studies have used K⁺ channel blockers to define operationally conductive pathways and to fingerprint the channels underlying the macroscopic conductance pathways. The relative contribution of K⁺ conductance pathways in determining V_a in the rabbit cortical collecting tubule (Frindt & Palmer, 1987) and Necturus gallbladder epithelium (Segal & Reuss, 1990), and E_m in rat colonic crypts (Burckhardt & Gogelein, 1992) have been established by the strong depolarisations evoked by K⁺ channel inhibitors. Large conductance voltage and Ca²⁺-activated K⁺ channels (maxi-K⁺) are found in many fluid-transporting epithelia and have been proposed to participate in the mechanism of electrolyte secretion in response to Ca²⁺-mobilising secretagogues (Petersen, 1986). The maxi-K⁺ channel is known to be blocked extracellularly by TEA and quinine at submillimolar concentrations and by Ba²⁺ at mM concentrations in Necturus gallbladder epithelium (Segal & Reuss, 1990). We can conclude by the absence of depolarisations induced in the presence of 5mM TEA that maxi-K⁺ channels, which contribute to the resting K⁺ conductance in Necturus gallbladder epithelium (Segal & Reuss, 1990), do not appear to contribute significantly to the maintenance of the resting K⁺ conductance in the crypt basolateral membrane.

Strong depolarisations were evoked by both 5mM Ba²⁺ and 20mM TEA in the ground region of unstimulated rat distal colonic crypts using the nystatin perforated-patch technique (Burckhardt & Gogelein, 1992). Thus the nature of the basolateral K⁺ conductance responsible for the maintenance of the resting membrane potential in the ground region of unstimulated rat colonic crypts appears to differ from that of the mid-region of crypts isolated from guinea-pig small intestine.

Hyperpolarisations induced by low concentrations of quinine (1-10 m M) may be consistent with the activation of a membrane K⁺ conductance, as has been previously reported to occur in Necturus gallbladder epithelial cells (Segal & Reuss, 1990). At low concentrations quinine, which is known to be highly membrane permeable (Iwatsuki & Petersen, 1985), induces the release of Ca²⁺ from intracellular stores (Isaacson & Sandow, 1967) which may explain the quinine-evoked hyperpolarisation if Ca²⁺-activated K⁺ channels are present in the crypt basolateral membrane. Alternatively the hyperpolarisation may be due to the inhibition of an apical Cl^- or Na^- conductance.

Outwardly-rectifying whole-cell K⁺ currents that exhibit time-dependent activation at depolarised potentials have been recorded in pancreatic acinar cells (Iwatsuki & Petersen, 1985) and in isolated Necturus enterocytes (Sheppard et al., 1991) using KCl rich pipette solutions, although the outward rectification observed may in part be accounted for by Goldman rectification due to asymmetry of K⁺ concentration across the plasma membrane. However the strongly outwardly-rectifying whole-cell currents observed in guinea-pig crypt enterocytes may be in part due to the presence of a non K⁺-selective conductance pathway, such as the outwardly-rectifying Cl^- conductance reported in guinea-pig villus enterocytes (Sepulveda et al., 1991). The identity of the single channels underlying the macroscopic outwardly-rectifying K⁺ conductance remain to be elucidated. Single channel recordings from excised patches of rat small intestinal crypt basolateral membrane have revealed the presence of large conductance Ca²⁺- and voltage-independent K⁺ channels that are blocked by intracellular Ba²⁺ (Fraser et al., 1991), whilst small conductance K⁺ channels that are Ca²⁺-dependent have been observed in cell-attached patches in isolated rabbit enterocytes (Sepulveda & Mason, 1985). Studies at the single-channel level will be required to determine whether either of these two channel activities contributes to the basolateral K⁺ conductance present in the unstimulated small intestinal crypt.

Return to top
References
Return to table of contents
Return to science homepage
Next chapter

Table 3.1. Composition of the solutions used in electrophysiological experiments.

Pipette free Ca²⁺ was buffered to 93nM in some experiments. All solutions were adjusted to pH 7.2 with Tris and the N-methyl D-glucamine salts (NMG) were prepared by titrating the appropriate acid. No correction for Ca²⁺ chelation by gluconate was made in preparing the low (8 mM) Cl^- Hanks medium. The 0 Ca²⁺ Hanks solution was prepared by omitting CaCl₂ from normal Hanks solution.